Delivery Model / Culture Handbook

Delivery Model and Culture Handbook

Collaboration · Cultivation · Delivery

We want a culture of collaboration and cultivation. We build teams that understand the purpose of their work, feel trusted to make decisions, and follow through together. We want useful improvements to reach customers sooner, with the quality they need and opportunities to learn from real use. This handbook describes the everyday choices that make that possible.

Collaboration means solving problems together across roles and Teams, sharing context, and taking responsibility for the overall result. Cultivation means developing people's judgment, skills, and confidence through coaching, experimentation, feedback, and opportunities to take greater responsibility. Leaders provide guidance and support, then step back as teams become more capable and confident.

Use this handbook when a decision is stuck, a commitment is threatened, two Teams need to work together, or an experiment produces an unexpected result. Everyone shares responsibility for this culture.

These are proposed working agreements for us to discuss and adopt together. They guide delivery and collaboration; company policies and established professional responsibilities still apply. We will make decision authority explicit before expecting people to exercise it.

How to use this handbook

Read it once to understand how we work, then return to the relevant chapter when you need it. Discuss the examples with your Team. If our everyday behavior contradicts what is written here, that is something to address together.

The culture we build every day

Culture is the accumulated result of what the organization repeatedly rewards, tolerates, punishes, and makes easy or difficult. A statement about trust means little if a reasonable decision brings punishment. A statement about ownership means little if every choice still needs a manager's permission.

Our culture becomes visible when work is difficult. When someone surfaces a risk, we help them examine it. When priorities conflict, we make a choice. When someone contributes a better idea, we change our thinking. Those responses teach people what is safe and expected here.

What we valueWhat you should observe
OwnershipPeople understand the intended outcome, make decisions within their authority, and follow work through to its effect on customers.
TrustManagers make decision boundaries clear and support people who act responsibly within them.
TransparencyRisks, mistakes, uncertainty, and changed assumptions become visible early enough to act.
LearningPeople test assumptions, inspect results, and change their approach when evidence warrants it.
CollaborationProduct, engineering, QA, security, operations, and customer-facing colleagues solve problems together.
QualityTeams address defects and technical risks where they arise and make unresolved risks explicit.
ImprovementPeople have time and support to improve the system they work in.
Customer focusWe check whether customers benefit from what we deliver and use that evidence to guide the next decision.

Principles that guide our work

These principles draw on Scrum, Lean, and Extreme Programming (XP) and guide how we put our values into practice.

PrincipleWhat it means in our work
Start with customer valueUnderstand the problem and the improvement we seek. Use customer and business evidence to judge whether the work helped.
Learn through small steps and rapid feedbackMake assumptions explicit, try something small, inspect the result, and adapt. Use building and testing to resolve uncertainty.
Choose the simplest solution that meets the current needAddress real requirements and risks. Let evidence justify additional complexity as the product evolves.
Improve the whole systemFollow work across Team boundaries through customer use. Judge improvements by their overall effect, including downstream delays and rework.
Create flow and pull work when capacity is availableLimit work in progress, reduce waiting, and help finish existing work before starting more. Make bottlenecks visible and address them together.
Build quality into every incrementKeep changes testable, integrate frequently, and improve the code as it evolves. Reduce scope when necessary while preserving the quality required for use.
Give teams the context and authority to actMake goals, constraints, and decision boundaries clear. Teams choose their approach, coordinate directly, and take responsibility for adapting and following through.

What we recognize

We recognize useful progress, early risk reporting, sound decisions, help given across teams, and work that prevents future problems. We celebrate small wins so people can see their contribution making a difference. A developer who helps another Team solve a problem or finish important work contributes to our shared success.

When the words and behavior disagree

Suppose we praise collaboration but criticize a developer for spending time helping another Team. The developer has learned what we actually reward. The manager's responsibility is to resolve that contradiction, make the priority explicit, and recognize the contribution. Anyone can raise these concerns, including people who are new here.

Start with the customer problem

Before we invest in a solution, we need a shared understanding of the problem and why it matters. A request for a feature is a starting point for a conversation. Product brings customer and business context; developers bring knowledge of what is possible and what it will take. Each needs the other's judgment.

The context every Team needs

Who is experiencing the problem, and what are they trying to accomplish?
What happens today, and why is changing it worth our attention now?
What improvement do we expect, and what evidence would help us judge it?
What constraints and obligations must the solution respect?

Product leads product direction, priority, and the definition of expected value. Developers help discover options, expose assumptions, and understand feasibility. Delivery leadership helps turn that understanding into a credible plan and makes obstacles visible.

How quarterly objectives help

An objective gives us a shared direction. Key results describe evidence of progress toward it. Completing an epic can be a useful milestone, but it does not by itself establish that a customer problem improved. Where a result depends on factors outside a Team's control, we acknowledge that and identify what the Team can influence.

For example, an objective might be to make a customer workflow easier to complete correctly. A smaller release could improve one difficult step. Product and the Team would agree what to observe, such as errors or successful completion, establish the starting point where possible, and allow enough time to learn from use.

When the connection is unclear

Ask for the missing context before treating the solution as fixed. Bring a possible interpretation or alternative so the conversation can move forward. Product should help resolve the question promptly; a Team should not have to invent business priorities to keep moving.

Some work protects existing service, reduces technical risk, or makes future changes possible. Explain that value in its own terms. Maintenance and migration work deserve an explicit place in priorities even when their benefit is less visible than a new feature.

How priorities change

Anyone can propose a better opportunity. Changing the agreed priority requires a visible decision by the people responsible for it, with affected teams involved. We explain what will move, what will wait, and what the change means for existing commitments.

Make decisions close to the work

We expect people closest to the work to make decisions within clear boundaries. That requires enough context to judge the options, access to the people affected, and clarity about who decides. A request to "take ownership" is incomplete if people cannot tell what they are authorized to do.

A proposed starting point for decision authority

Agree named decision owners using these proposed boundaries. Resolve differences with existing authority before adopting them.

DecisionWho leads itWho must be involved
Product priority and intended outcomesProduct within business directionEngineering and delivery assess feasibility, capacity, risk, and dependencies.
Implementation within agreed standardsDevelopers in the TeamConsult relevant engineering, QA, security, and operations expertise as the risk requires.
Shared architecture or data contractsDesignated engineering ownerAffected Teams and service owners review compatibility, migration, and operational impact.
Scope, dates, or conflicting commitmentsAccountable product or business ownerTeams and delivery provide options; leadership resolves conflicts beyond delegated authority.
Production readiness and customer releaseEngineering or operations for readiness; Product for availabilityQA, support, implementation, and other required owners coordinate.

Decide without waiting for certainty

For a reversible decision within your authority, seek input suited to the risk, make the decision, and state what would cause you to revisit it. Decisions with significant customer, data, security, or operational consequences deserve stronger evidence and the appropriate review. The cost of delay belongs in the discussion alongside the cost of being wrong.

Once we understand the current need, have a workable approach, and have addressed the risks that matter before starting, we make a decision and begin building. We do not need agreement on a perfect solution or answers to every future question. When uncertainty remains, choose a small implementation, prototype, or experiment that can help resolve it. Agree what we need to learn and when we will review the result. Revisit the decision when new evidence or a material risk warrants it; the possibility of a more elegant solution alone should not keep restarting the discussion.

Record the decision, the reason, the owner, and any important assumptions where affected people can find them. When discussion repeats, ask what new evidence or unresolved risk would change the choice. If authority is unclear or disagreement cannot be resolved in time, bring options and a recommendation to the appropriate decision owner.

A shared interface is a shared decision

A Team may choose how to implement its own component. Changing an interface another Team relies on requires coordination with that Team and the designated technical owner. Autonomy includes responsibility for the effects of our decisions on others.

Deliver something useful and improve it

We want to get something good into customers' hands, learn from it, and improve it. "Good" means useful for the intended need and fit for safe, reliable use. That defines what the next release needs to accomplish.

Separate today's need from tomorrow's possibilities

Developers should challenge assumptions about future scale, flexibility, and features. Some decisions need investment now because changing them later would be costly or dangerous. Others can wait for evidence. Make that distinction explicit rather than treating every possible future need as part of the first version.

What must a customer be able to accomplish in this release?
What quality standards and operational needs must this release meet?
Which decisions are expensive to reverse, and which can evolve?
What can we leave out while preserving a useful, coherent result?
What do we expect to learn, and when will we review it?

Build in slices that can teach us something

Work with Product to identify a small part of the customer workflow we can make usable. Include the testing, integration, and operational work needed for that path. A collection of unfinished technical layers may be necessary progress, but we should understand what still prevents it from reaching a user or testing an important assumption.

For example, a Team considering a flexible workflow engine could first support one important customer workflow. It would preserve required controls and data correctness, while postponing speculative configuration options. If shared architecture requires a broader foundation first, explain the dependency and test the riskiest assumptions early. We deliver in small increments while accounting for technical constraints.

Learn from each release

Before release, agree who will gather feedback and what will happen to it. After release, review how customers use the change, what they tell us, and how the service performs. Product and the Team then decide whether to expand, adjust, or stop.

Protect teams from moving expectations

Managers support the agreed scope boundary. A team should not be encouraged to deliver a modest first version and then criticized for omitting features everyone agreed to defer. If new information changes the need, acknowledge the change and make a new decision.

Celebrate a useful improvement and the judgment behind it. Plan time for the next iteration so we can follow through on improvements we have agreed to make.

Take ownership across Team boundaries

A Team is a team that works together toward shared goals, building continuity, shared understanding, and trust. Its responsibility includes collaborating with other Teams to deliver useful improvements to customers.

Follow the work beyond your own contribution

Ownership means knowing the intended result, making your part visible, and following through on the agreements you make. You do not have to perform every task yourself. You do need to make sure the next person understands what they are receiving and that unresolved issues have an owner.

When work is waiting for review, testing, an environment, a decision, or another team, make the waiting visible. Ask whether helping finish that work is more valuable than starting something else. A busy Team can still be contributing to a slow delivery system.

Work directly with the people involved

If another Team's work affects yours, start the conversation with them. Explain the dependency, the needed timing, and the effect on the shared result. Agree who will do what and by when, and capture the agreement where both teams can see it. Include people who will need to support or operate the result.

Delivery leadership can facilitate the first conversations, help resolve disagreement, and follow through when coordination is failing. As those habits develop, direct collaboration should become the default. The delivery lead remains available when teams need help.

When another Team cannot meet the need

Suppose your work depends on a shared API change that the other Team cannot fit into its current priorities. A request in their queue is not an accepted commitment. Talk through a smaller interface change, a different sequence, or a safe temporary approach. If the conflict still needs a business priority decision, escalate it together with the options and their consequences.

For each significant dependency, record which team is providing the work, who needs it, what they have agreed to deliver, the needed date, and the next action. Keep that record lightweight and current. Escalate early enough to give people time to act.

Improve the whole path

The wider delivery group exists to solve problems that a Team cannot solve alone: shared capacity, integration, dependencies, and conflicting priorities. If one Team improves its local numbers by pushing unfinished work onto another, the system has not improved. Leaders should look for where work accumulates and help teams relieve that constraint.

Make commitments people can rely on

A useful commitment makes clear what people are relying on and what each of us has agreed to do. We take those agreements seriously while acknowledging that software work contains uncertainty. Credibility grows when we make informed commitments, act on risks, and speak honestly when the facts change.

Be clear about what kind of agreement we are making

A business target states an outcome or date we want to achieve. A forecast describes what the available evidence suggests is likely. A commitment states an agreement people are expected to act on. These can inform one another, but calling a target a commitment does not make the work feasible.

Before making an external promise, the accountable owner needs input from the people doing the work. Record the intended result, what completion means, relevant assumptions and dependencies, and how changes will be handled. If a date is fixed, make scope and sequencing choices early. If scope is fixed too, leadership must address the resulting capacity and risk honestly.

Shape the plan together

Understand the priority, customer need, and minimum useful result.
Include support work, leave, maintenance, quality, and integration in capacity discussions.
Check material dependencies with the teams expected to provide them.
Expose uncertain work and choose what to investigate before relying on it.
Agree how progress and threats will be reviewed, including who can authorize changes.

When a commitment is threatened

Raise the issue when the evidence changes, even if you do not yet have the answer. Explain the original agreement, what changed, the likely effect, and the next decision needed. Bring recovery options where possible: reduce scope, change the approach, resequence work, resolve a dependency, or revise the date with the accountable owner.

Continue to act while the decision is being resolved. Keep affected people informed, and update the shared record once a change is agreed. Do not silently redefine the original promise or move the date in a tracker and assume the people relying on it will notice.

Accountability works in both directions

Teams are responsible for preparation, follow-through, and timely risk reporting. Managers and Product are responsible for honoring priorities, making timely decisions, and recognizing the consequences of new requests. If leadership adds urgent work, it must help decide what gives way.

After a miss, examine the causes and our response. An honest forecast can prove wrong; an emerging problem can also be ignored. Those situations require different action. Better reliability comes from understanding both the delivery conditions and the choices people made.

Use a cadence that helps us learn

We use quarterly planning to set direction, Sprints to learn and adapt, and regular conversations across Teams to coordinate the work. We will adjust this cadence as we learn what helps teams make decisions and work together.

Quarterly direction and planning

Product and leadership establish the direction and explain the intended outcomes. Before planning, Product, engineering, and delivery work through readiness, major unknowns, capacity, and dependencies. Teams help shape a feasible plan. The result should make priorities, important agreements, and unresolved risks visible, including the work we are choosing to defer.

When several Teams plan together, give them shared context, time to develop their plans, and opportunities to resolve dependencies and review risks. Use enough structure to support that coordination. A quarterly plan remains open to new evidence throughout the quarter.

Sprint learning within each Team

Where we use Scrum, the Sprint gives us a short cycle for pursuing a meaningful goal, inspecting what we create, and adapting. At planning, the Team understands the purpose and selects a realistic approach. During the Sprint, developers coordinate and adjust as they learn. The daily conversation should help them move toward the goal.

At the review, inspect the actual product with relevant stakeholders and discuss what the evidence means for the next step. The retrospective examines how we worked and selects an improvement worth trying. Scrum supports empirical learning; we use additional delivery practices where we need to understand flow, forecasts, and release coordination.

Coordination across Teams

Use a weekly Sync to align Product, engineering, delivery leadership, and Team representatives on priorities, dependencies, integration, risks, and decisions. Participants leave with clear actions, owners, and timing. Routine status can be shared in a concise written update; a material problem should not wait for the next meeting.

Release and customer feedback

Teams coordinate readiness, operational support, customer communication, and the decision to make a change available. Release when the agreed conditions are met; a Sprint boundary alone does not determine release timing. Feedback from customers and operations feeds the next product and team decisions.

Improve the cadence itself

Ask periodically what each meeting helps us decide or learn. Shorten, combine, or replace meetings that no longer help, and agree how any necessary decisions or coordination will happen. If urgent support work needs a different approach, agree how to handle it and keep its priority and progress visible.

Learn openly and follow through

People should be able to ask for help, challenge an idea, report a mistake, or admit uncertainty without humiliation or retaliation. That safety allows us to see what is actually happening. We also expect people to act responsibly, honor agreements, and learn from the consequences of their choices.

Agree how to run and review experiments

Before starting an experiment, agree what you want to learn and how you will test it. Set limits on cost and risk, decide what evidence to collect and when to review it, and name who will run it. Involve the people affected and get the approval needed for the risks involved.

Suppose a Team tests whether an earlier design conversation reduces rework. The team agrees to try it for two Sprints and inspect the result. If rework does not improve, that can still be a well-run experiment. We examine the evidence and choose whether to change or stop it. If nobody holds the conversation or gathers the agreed evidence, we have a follow-through problem to understand.

Respond to a mistake with investigation

First contain the harm and support the people addressing it. Then reconstruct what happened using the information available at the time. Ask what made the choice appear reasonable, which controls or signals were missing, and what would help us avoid the same problem. A bad outcome alone does not establish careless behavior.

Give feedback about observable behavior and its effect. "The dependency change was known on Tuesday but was not shared until Friday, leaving the other Team unable to adjust" gives us something to examine.

When coaching and expectations are not enough

Discuss the pattern privately with the person. Check whether the agreement was clear, achievable, within their control, and supported by the necessary skills and authority. Listen for conflicting demands or other obstacles. Agree on concrete changes, the support to provide, and when to review progress.

If the pattern continues, involve the person's manager with specific evidence of the behavior, its delivery impact, and the support already provided. Work together to determine what needs to happen next.

Leaders make speaking up safe

When someone raises a problem, thank them for speaking up and work with them on what to do next. Ask questions to understand what happened and give them room to explain. Share your own mistakes and what you changed as a result.

Build quality into the work

Quality is part of delivering something useful. A smaller release still needs to protect customer data, behave correctly for its intended use, and be supportable in operation. We reduce unnecessary scope and avoid speculative work while keeping the controls that the actual risk requires.

Agree what ready for use means

Each Team needs a shared Definition of Done that reflects its product and the organization's standards. Include the relevant review, testing, integration, security, and operational expectations. Make any remaining release conditions explicit so finished development is not mistaken for customer availability.

Developers and QA work together from the beginning to understand behavior and risk. Automated checks should provide timely feedback where they are useful; exploratory work and specialist review address questions automation cannot settle alone. Infrastructure, operations, and security expertise should enter early enough to influence the design.

Follow the change into production

Agree how the team will detect problems after deployment and who responds. Match the rollout and recovery plan to the risk, including rollback or another safe recovery option where practical. Coordinate customer communication and support readiness before making the feature available.

Be clear about how far the work has progressed. Code merged, deployed to production, available to users, and producing an expected benefit are different milestones. A feature can be deployed behind a toggle while customer release waits for readiness. Our reporting should make clear which milestone we have reached.

Treat migration as a customer and operational change

During a move from legacy platforms to a shared platform and data architecture, a change may affect multiple systems and teams. Make interface ownership, data reconciliation, cutover conditions, support coverage, and recovery plans explicit. Include the ongoing needs of legacy services in capacity rather than assuming they will take care of themselves.

For example, before releasing one part of a migration, we may need a way to check its data and monitor how it affects operations. A feature demonstration is not enough evidence to retire the previous path. The technical and business owners need to agree what evidence is required before making the transition.

Make technical debt a visible choice

When a short-term approach creates future cost or risk, explain the consequence in language Product and leadership can use. Record who is responsible, why the choice is justified, and when it will be revisited. Prioritize repair according to its effects on reliability, change effort, customer needs, and risk.

Plan time for service health and improvement. Use recurring incidents, unreliable tests, and manual release work to identify where that effort is needed. Account for this work when deciding how much new development the team can take on.

Use evidence with understanding

Evidence should help us understand the work and make better decisions. A number without its definition, context, and limitations can lead us in the wrong direction. We examine measures with the people doing the work and connect what we observe to a question worth answering.

QuestionEvidence that can helpHow we interpret it
Are customers benefiting?Customer feedback, adoption, task success, errors, or other agreed outcome measuresProduct leads interpretation. Delivery of a feature alone does not prove benefit.
Are we meeting agreements?Epic or milestone completion against an explicit baseline; Sprint Goal results; changes to scope and datesKeep original agreements and approved changes visible. Distinguish completion from value.
Where is work getting stuck?Work age, cycle time, throughput, work in progress, blockers, and dependency waitLook for patterns, comparable work, and causes of delay before choosing an intervention.
Is the product healthy?Defects, incidents, failed changes, recovery, and relevant service measuresConsider customer impact and reliability alongside delivery speed.
Is the system sustainable?Unplanned work, recurring overtime, support demand, and team feedbackInvestigate capacity, skills, priority changes, and working conditions.

Protect the meaning of a measure

We do not rank developers by ticket counts, story points, or lines of code. We do not compare Team velocity as if each team's estimates were a common unit. Teams choose the estimation method that works for them. Anyone using those estimates should understand the assumptions and uncertainty behind them. Watch for incentives that reward smaller promises, inflated estimates, or hiding unfinished work.

For an improvement target, establish the starting point, what is counted, the time period, and how changes will be treated. A reported improvement in delivery against commitments means little if the definition changes along the way. Review trends and variation before attributing one good or bad period to an intervention.

What leadership should be able to understand

Are we on track?
Are we achieving the expected value?
What could prevent success?
What decisions or support are needed?

A concise four-blocker update can organize completed work, work in progress, upcoming work, and action items or issues. Include the evidence, uncertainty, and decisions needed to answer those four questions. Combine the relevant perspectives from Product and engineering; delivery reporting alone cannot establish business value.

Lead in ways that build capable teams

Managers create the conditions in which ownership can work. That means giving people direction, enough context to make decisions, clear boundaries, and timely help. It also means changing management habits that quietly teach people to wait for permission or hide unwelcome information.

Give direction and support judgment

Explain the intended result and the constraints. Ask people for their recommendation and reasoning, then let them act within the authority you have agreed. If you override a decision, explain why and whether the boundary has changed. Repeated unexplained overrides will teach the team that delegated authority is unreliable.

Provide more direct support where a person or team needs it. Facilitate difficult conversations, model how to work through dependencies, and help them test their thinking. Step back as people gain the experience and confidence to make those decisions themselves.

Remove obstacles you are responsible for

Resolve conflicting priorities, make decisions in time for teams to act, and address capacity or skill gaps. When everything is urgent, choose what will wait. When teams repeatedly depend on one specialist, help reduce that bottleneck through pairing, documentation, or appropriate staffing. Protect room for maintenance, improvement, and learning.

Repeated overtime is a reason to examine the plan and working conditions. An exceptional recovery effort may be necessary, but it should include recovery time and changes that reduce recurrence. Recognize the person who prevents an emergency as well as the person who helps resolve one.

Use coaching and feedback throughout the work

For developers, recognize technical judgment, maintainability, collaboration, mentoring, customer understanding, and responsible follow-through. Invite feedback from colleagues who see the work, using concrete examples. Make growth expectations explicit and support them with opportunities to practice.

Help people contribute beyond their specialty

Encourage developers to understand customer workflows and operational realities, and help managers and Product understand technical constraints. Pair people across specialties when it advances real work. A person can deepen their expertise while learning enough about adjacent work to collaborate effectively.

Leaders help build enthusiasm by connecting developers to the purpose of the work and giving them a meaningful voice in shaping it. People have more reason to feel invested when they understand the customer's problem, their ideas influence the solution, and they can see the difference their work makes. Recognizing useful progress, creating opportunities to learn, and helping teams overcome obstacles reinforce that investment. These everyday actions help teams build confidence and take pride in what they accomplish together.

Get started and keep improving

We begin with the teams and work we have. Understand what already works before changing it, and involve people in shaping the next step. The handbook gives us a shared direction; experience tells us where we need to clarify or improve it.

Understand the system you work in

Learn how customers use the workflows your Team supports and which problems the Team is trying to solve. Work with Product and customer-facing colleagues to keep that understanding current.
Follow work through development, testing, integration, release, and feedback. Look for where it waits, where handoffs create friction, and where ownership becomes unclear.
Identify where your Team depends on other teams and where they depend on you. Communicate directly with the people involved, make dependencies explicit, and keep shared agreements accessible and current.
Be clear about the decisions you can make, the quality expectations you must meet, and how to raise risks. Work with your manager and the relevant decision owners to resolve unclear or conflicting boundaries.
Work with colleagues across roles to understand how your decisions affect the wider delivery system. Use what you learn to identify and test improvements together.

What each Team should agree together

Keep a short, accessible working agreement covering the Team's purpose, decision boundaries, Definition of Done, coordination habits, and how it handles urgent work. Identify where commitments, dependencies, and decisions are visible. Agree how to raise a concern when speaking directly to the person involved is difficult.

The first improvement should address a real problem the team recognizes. Establish what happens today, choose a small change, and set a review point. A Team waiting days for clarification might try a direct conversation with Product when a question arises and inspect whether that reduces waiting without creating constant interruption.

When our practice falls short

If people still wait for permission, check whether they know which decisions they can make and trust that leaders will support them. If commitments repeatedly slip, examine readiness, competing work, dependencies, and follow-through. If meetings grow while decisions remain unresolved, clarify who decides. If the dashboard looks better but customers see no improvement, revisit what the measures mean.

Anyone can propose a handbook change with a concrete example and a suggested improvement. Delivery leadership should help bring affected people together, confirm the decision with the appropriate owner, and publish the revised agreement. Review the handbook during quarterly reflection and when important changes expose a gap.

The next conversation

Choose one recurring situation in your Team. Compare what happens today with the approach described here, and discuss what support would help you change it. Agree on one action and a time to review how it went.

Discovery from ideas to evidence

Use a shared Portfolio Kanban view to track ideas from initial consideration through delivery and review of customer results. Product leads discovery with engineering and the people who understand the customer problem.

Handbook diagram 1: Discovery from ideas to evidence

At each decision: continue, adjust, defer, or stop. Limit work in progress, name the next decision owner, and pull prioritized work when capacity is available. Discovery continues during delivery.

ConversationWhenResult
Intake and prioritizationWeeklyProduct, engineering, and relevant business owners decide what to explore, prioritize, defer, or stop.
Discovery working sessionAs neededProduct and relevant specialists test an assumption through research, a prototype, or a small implementation.
Outcome reviewMonthly or at the agreed evaluation pointProduct, the Team, and business owners inspect customer and operational evidence and decide the next investment.

Delivery cadence at a glance

Starting cadence: quarterly direction, two-week Sprints, and a weekly ART Sync. Adjust together as we learn what the work requires.

Handbook diagram 2: Delivery cadence at a glance

ART Sync brings Product, engineering, delivery leadership, and Team representatives together to resolve cross-Team priorities, dependencies, integration, risks, and decisions.

ConversationWhenResult
Quarterly planningEach quarterProduct, leadership, and Teams agree objectives, a feasible plan, capacity tradeoffs, and dependencies.
ART SyncWeeklyAgreed cross-Team actions, owners, dates, and decisions requiring leadership support.
Objectives reviewMidquarterProduct, leadership, and Teams inspect results and adjust priorities or commitments.
Leadership updateWeeklyDelivery, Product, and engineering share progress, value evidence, risks, and decisions needed.
Quarterly reflectionEnd of quarterTeams and leaders choose improvements to the wider delivery system.
Release readinessAs neededProduct, engineering, QA, operations, and support resolve readiness and rollout decisions.

Scrum as a learning cycle

Each Sprint pursues a goal, creates useful increments, and provides evidence for the next decision. The Team adapts its plan as it learns.

Handbook diagram 3: Scrum as a learning cycle
ConversationParticipantsStarting cadence
Sprint PlanningScrum TeamStart of each Sprint
Daily ScrumDevelopersEvery working day
Sprint ReviewScrum Team and stakeholdersNear Sprint end
RetrospectiveScrum TeamAfter Review
Backlog refinementProduct Owner and relevant DevelopersWeekly (or focused sessions as needed)

Refinement maintains shared understanding of upcoming work throughout the cycle. A usable increment meets the Definition of Done.